pa-skills

Knowledge graph and smart memory management using graphify + Obsidian-inspired patterns. Use when: setting up a knowledge graph, managing memory health, cross-linking notes, compiling wiki pages from scattered notes, adding structured frontmatter, or running memory health checks. Triggers on: 'knowledge graph', 'graphify', 'wiki', 'cross-link', 'memory health', 'frontmatter', 'compile notes', 'wikilinks'.

AI Personal Assistant network skill. Use when: contacting another PA, coordinating with peer agents, scheduling meetings between owners, broadcasting messages, looking up contacts from the local PA directory, or performing any cross-PA task. Reads contact data from data/pa-directory.json in the workspace. Does NOT contain hardcoded contact data — each PA maintains their own directory file.

Route and execute inter-PA coordination tasks in the OpenClaw multi-agent PA network. Triggers when: contacting another PA, passing a message between owners, broadcasting to a PA group, handling group lead/silence rules, or adding a new PA to the directory. NOT for: contact lookups (use contact-list skill), calendar invites (use meetings skill), or billing issues (use billing-monitor skill).

Automatically creates new skills from complex problem-solving sessions. Triggers at the end of multi-step tasks that required novel approaches, debugging sessions, or workflow creation. Use when: a task took 3+ tool calls, involved a non-obvious solution, or established a pattern likely to recur. Also triggers on: "create a skill from this", "save this as a skill", "this should be a skill".

Step-by-step setup and troubleshooting for connecting an owner's Google Calendar to an OpenClaw agent with read and write access. Use when: setting up calendar access for the first time, fixing read-only access, re-authenticating a broken Google connection, explaining the difference between the owner's Google account and the agent's Google account, how an OpenClaw agent accesses Gmail or Google Workspace, where credentials are stored, how to use the agent inbox, how to give the agent write access to the owner's calendar, or guiding another agent or user through the email or calendar setup model.

Search past WhatsApp/chat conversations stored in the audit log PostgreSQL database. Use when the user asks about past conversations, what was discussed, what someone said, finding a specific message, or referencing previous discussions. Also use to reply to or quote specific past messages.

Enforce immediate execution of any commitment made in a reply. Use this skill before finalizing ANY reply that contains a promise to act — phrases like "אדווח", "אעדכן", "אשלח", "אחזור אליך", "I'll send", "I'll report", "I'll follow up", "I'll update". Prevents the agent from saying it will do something and then not doing it.

Single source of truth for all contacts, PAs, and WhatsApp groups. Use when: looking up a phone number, email, address, PA, group JID, adding/updating a contact, or before sending any WhatsApp message. Reads and writes contact-list.md in the workspace root.

4-layer recall procedure (durable → semantic → PostgreSQL → daily notes) before claiming "no context". Invoke when user explicitly asks "do you remember", "what did we discuss about X", "search history". NOT a general assistant — only fires when recall is the explicit task. Drop the prior "use proactively" rule. Triggers: "do you remember", "what did we discuss", "when did", "find that conversation about".

Evaluate everything the PA agent manages — tasks, skills, PA network health, billing, calendar connections, and memory quality. Use when: owner asks for an evaluation, wants to know what's working and what isn't, or requests a performance report. Combines supervisor status with quality scoring.

Complete WhatsApp management for OpenClaw agents: per-conversation memory (groups + DMs), unanswered message tracking, loop prevention, and multi-PA coordination. Use when: tracking conversation context, recalling past decisions, finding unanswered messages, or preventing echo/duplicate message loops.

Runbook skill for failures: cron error, PA failure, cascade, gateway disconnect, semantic DB stale. Walks: detect → classify → diagnose → notify → log. Replaces ad-hoc failure handling. Triggers: "cron failed", "X is broken", "cascade", "incident", "gateway down", "PA failure".

Schedule meetings, parse meeting notes, extract action items. Three distinct intents covered. Invoke when user mentions a specific meeting (past or future) by name or context. NOT for general calendar checks (use direct calendar.py), NOT for time-zone math. Triggers: "schedule with X", "minutes of the X meeting", "action items from yesterday's meeting", "prep me for meeting with X".

Create and troubleshoot Ocana agents. Use when creating a new Ocana agent, refreshing/repairing management token auth, diagnosing `ocana_agent_create` failures, checking `gatewayReachable`, handling agents stuck in `provisioning`, debugging `226/NAMESPACE` or missing `/usr/bin/openclaw`, or testing lifecycle/reprovision flows for a newly created agent.

Explain how email and Google Calendar work for OpenClaw agents, including the distinction between agent email and owner email, how gog and the Google Workspace skill are used, where related credentials/config live, and how to troubleshoot common setup or access issues. Use when someone asks how an OpenClaw agent accesses Gmail/Google Workspace/Calendar, where credentials are stored, how to use the agent inbox, how to give the agent write access to the owner's calendar, or how to guide another agent/user through the email or calendar setup model.

Generate and send a daily briefing to your owner covering today's meetings, urgent emails, open tasks, and anything that needs attention. Use when: it's the start of the owner's day, when asked for a summary, or on a scheduled cron job.

Unified workspace maintenance skill. Covers: (1) backing up the workspace to GitHub git repo, and (2) updating OpenClaw and skills. Use when: asked to backup, push to git, update openclaw, update skills, or run maintenance.

Step-by-step onboarding guide for setting up a new AI Personal Assistant on OpenClaw. Use when: a new PA is being created, someone asks how to set up an agent, or guiding a user through the full setup process from account creation to first response.

Personal CRM built on monday.com. Tracks contacts, last interactions, next meetings, and topics. Runs daily to update from Calendar + email. Delivers pre-meeting briefings as part of morning-briefing. Use when: someone asks about a contact, before a meeting, or during daily CRM sync.

Configures autonomous heartbeat and cron-based proactive routines (e.g., morning briefing setup, unanswered-message scan cadence). Invoke when user is SETTING UP or MODIFYING a proactive routine, not when the routine itself fires. Cron crons execute their own logic — they do NOT load this skill. Triggers: "set up proactive check", "configure heartbeat", "add morning briefing", "tune cron cadence".

Build and run the owner's self-branding system in a structured way across LinkedIn, Substack, dev.to, X, profile rewrites, recurring series, and comment strategy. Use when the owner asks for positioning, a branding plan, profile updates, content drafts, a publishing sequence, comment ideas, or help maintaining a consistent public voice. Optimize for operator credibility: specific experience, defensible numbers, concrete artifacts, and category ownership.

Proactive self-monitoring of infrastructure, services, and health. Tracks disk/memory/load, service health, cron job status, recent errors. Auto-fixes safe issues. Triggers on: health check, heartbeat, monitor status, service status, infrastructure check.

Startup skill router: maps any incoming request to the correct skill using Quick Lookup table and _manifest.json. Loaded automatically at session start. Triggers on: every new request — this is the dispatch layer, not a user-facing skill. NOT for: executing tasks directly — always routes to a domain-specific skill.

Decide where to save any piece of information — monday.com, local file, daily notes, or MEMORY.md. Use this skill before saving anything to ensure the right destination. Prevents local clutter and ensures the owner can access all relevant content in monday.com.

Central status dashboard for the PA agent. Use when: owner asks 'what's the status', 'what are you working on', 'what's happening', or any status/overview question. Aggregates active tasks, open issues, monitored groups, pending follow-ups, and system health into one structured report.

Report AI token usage and estimated costs. Use when: owner asks about costs today/yesterday/this week, per session, or per model. Shows main session, cron jobs, and subagents. Answers: 'how much did today cost?', 'how much was this session?', 'what was last week's spend?'

Transcribe WhatsApp voice messages using local Whisper CLI. Use when: owner or contact sends an audio/ogg voice message. Combines Whisper transcription + CRM update + task creation. Works offline for short clips, uses OpenAI API for long clips. Hebrew and English supported.

OpenClaw API gateway proxy for authenticated SaaS calls (100+ services in references/SERVICES.md). Use ONLY when making an HTTP request to a third-party service that requires the gateway's stored credentials. NOT for direct API calls with bearer tokens already in scope, NOT for monday.com (use monday-for-agents), NOT for WhatsApp (use heleni-whatsapp). Triggers: "call the X API", "fetch from <service>", "send to Slack via gateway", "use the gateway for".

Turns data/commitments.jsonl (flat list) into a who-owes-what-to-whom graph. Surfaces overdue chains and cross-PA dependencies that the flat list hides. Triggers: "what's overdue", "what did X promise", "commitment graph", "who's blocking whom".

Structured build workflow: branch, delegate to coding agent, test, screenshot, and commit. Triggers on: "$devprocess", "build this properly", "use the coding agent workflow", or any feature/refactor/infrastructure change that is too large for a direct edit. NOT for: one-line fixes, doc edits, prompt tweaks, or simple config changes — do those directly.

Multi-phase dialectical stress-test for HIGH-STAKES decisions only (architecture choices, irreversible product calls, strategic bets). Heavy-cost skill — do NOT use for routine questions, brainstorming, or simple tradeoffs. User must explicitly invoke or describe a decision they call "high-stakes", "irreversible", or "needs stress-testing". Triggers: "stress test this decision", "dialectic on", "challenge this thesis", "should I really".

Scans HOT.md and proposes which rules should graduate to SOUL.md (untouched 30+ days) or be removed entirely. Run weekly via cron, or invoke ad-hoc when HOT.md feels stale. Triggers: "clean up HOT", "graduate rules", "is HOT.md stale", "what HOT rules can I drop".

Single-skill wrapper for the 4-layer recall (durable memory → semantic DB → PostgreSQL → daily notes). Use BEFORE saying "I don't have context". Triggers: "search memory", "what do we know about X", "deep recall", "I don't remember".

Multi-tiered memory management (HOT/WARM/COLD) for context compaction. Invoke ONLY for explicit compaction events: post-`/compact` cleanup, MEMORY.md tier promotion, archive batch, or "trim my context". NOT for general recall (use deep-recall) or routine memory writes (use storage-router). Triggers: "compact memory", "promote to durable", "archive old context", "tier this".

Set up a monday.com account for an OpenClaw agent and work with monday.com boards, items, and updates via the GraphQL API or MCP server. Use when: creating a monday.com workspace for a PA, connecting the PA to monday.com, querying boards and items, creating or updating items, troubleshooting monday.com API access, self-registering an agent on monday.com via HATCHA agent verification, or integrating with monday.com workflows. Covers GraphQL cookbook, column types, MCP configuration, and HATCHA self-registration. Works with any LLM model.

DEPRECATED — merged into monday-for-agents. See skills/monday-for-agents/SKILL.md.

Reduce OpenClaw token usage and API costs through smart model routing, heartbeat optimization, budget tracking, and native 2026.2.15 features (session pruning, bootstrap size limits, cache TTL alignment). Use when token costs are high, API rate limits are being hit, or hosting multiple agents at scale. The 4 executable scripts (context_optimizer, model_router, heartbeat_optimizer, token_tracker) are local-only — no network requests, no subprocess calls, no system modifications. Reference files (PROVIDERS.md, config-patches.json) document optional multi-provider strategies that require external API keys and network access if you choose to use them. See SECURITY.md for full breakdown.

Autonomous build-ship-verify loop that ends with code committed, deployed, and confirmed live. Triggers on: "$ownership", "execute X with ownership", "build this overnight", or any request for an autonomous product loop with no human checkpoints until production is verified. NOT for: one-off tasks, quick fixes, or supervised dev work (use devprocess for those).

Runs before every git commit on heleni-memory. Checks: no credentials in tracked files, no oversized files, valid SKILL.md frontmatter for any new/modified skill. Triggers: "commit", "before push", "validate repo", or auto-runs as a git pre-commit hook.

Turn any markdown into a shareable web page via md.page. Use when the user asks to "share this", "publish this markdown", "create a shareable link", "make this a web page", "send this as a link", "host this", or wants to turn any markdown content into a URL. Also triggers on "publish a report", "share my notes", "create a page", or any request to make content accessible via a link.